Zillow Premier Agent ROI: how to measure the commission Zillow leads actually close, not the connections you paid for

Zillow Premier Agent ROI is the commission from closings that started as Zillow connections, less what you paid for them, divided by what you paid. Measure it by the month each connection arrived and follow it for 12 to 18 months. This month's bill against this month's closings is the wrong comparison.

Zillow is the largest paid lead source most residential agents and teams will ever buy from, and it is the one they find hardest to judge. The invoice arrives every month. The closings arrive whenever buyers are ready, which can be next month or next year. In between sit dozens of connections, a few of whom were already in your phone, a few of whom also inquired on another portal, and most of whom never transact at all.

This post covers how Zillow's two pricing models charge, why the obvious ROI calculation misleads under both, a worked example measured three ways, the break-even between share of voice and a pay-at-close fee, and the routine that turns your CRM and your closed transactions into a number you can defend before you renew a zip code.

What is Zillow Premier Agent ROI?

Zillow Premier Agent ROI is the return on what you pay Zillow, measured in the commission you earn on transactions whose buyer or seller first came to you as a Zillow connection. It is a closed-commission measure, not a lead count, and it needs a time window long enough for those connections to close.

The formula is the ordinary one from how to calculate marketing ROI: commission earned minus cost, divided by cost. What makes Zillow hard is not the arithmetic but the two inputs. The commission has to come from closings you can tie to specific Zillow connections, matched on the client's phone number or email, and the cost has to be the cost of those same connections, not of whichever month the closing happened to land in.

Lead counts and cost per connection are useful for watching the program week to week, and Zillow reports them to you. They do not tell you whether the program pays. A zip code with cheap connections and no closings loses money; one with expensive connections that close at twice the rate may be your best spend. Cost per lead covers why the cheaper lead is so often the dearer customer.

How Zillow Premier Agent pricing works

Premier Agent comes in two pricing models. Market-based pricing is a monthly prepaid budget by zip code, with connections shared in proportion to your share of the spend. Zillow Preferred, formerly Flex, charges nothing upfront and takes a performance fee when a transaction closes with one of the leads.

Under market-based pricing, then, you are buying a share of a zip code's connections, not a fixed number of them. If another team raises its budget in the same zip, your share and your connections fall while your spend stays the same. Cost per connection is an output of the market, and it moves month to month without anything changing on your side.

That two-year horizon is the most useful number Zillow has published for anyone measuring ROI. It is Zillow telling its investors how long its own leads take to turn into transactions. If Zillow books revenue on a two-year view, an agent who judges the program on a one-month view is measuring something else.

Your agreement is the authority on your own fee and terms, and Zillow can change them. For measurement, the point is structural: under market-based pricing you pay for connections whether or not they close; under Preferred you pay only on closings, but on every one of them. The two models put the risk in different places, and they need the same measurement to compare.

Why the monthly Zillow bill and the closings never line up

A Zillow connection can take months to close, so the closings in any month come from connections bought across the past year or more. Dividing this month's commission by this month's spend credits old leads to new money and makes the program look better or worse depending only on timing.

Picture a team that raises its zip code budget in January. For the first quarter the spend is higher and the closings are those of last year's smaller budget, so the program appears to have collapsed. By autumn the January connections start closing and the program appears to have recovered, even if nothing about lead quality changed. The same happens in reverse after a budget cut, which looks like a triumph for two quarters. Long sales cycles covers this lag in general; residential real estate is one of its clearest cases.

• Spend is monthly and immediate. Market-based pricing is prepaid, so the cost of a connection is fixed the day it arrives.
• Commission is lumpy and late. It lands at closing, after a search, an offer, inspections and financing, and only for the minority of connections who transact.
• The calendar mixes cohorts. This month's closings include connections from many past months, each bought at a different share of voice.
• Seasonality compounds it. Spring searchers inquire early and close in summer; a calendar view of the same year credits the summer with spring's spend. Seasonal attribution works through the same effect for other businesses.

The fix is to stop asking what Zillow earned this month and ask what each month's connections earned eventually. Group connections by the month they arrived, match each to your closings, and let the cohort's commission accumulate until it stops growing. That is the number to set against that month's spend.

Worked example: one year of Zillow spend, measured three ways

The same year of Zillow connections can show a 17% return, a 163% return or a 104% return, depending on whether you count calendar closings, every matched closing over 18 months, or only closings whose earliest source was Zillow. Only the last is the program's real contribution.

Here is an invented two-agent buyer team in one metro. It spent $3,000 a month in market-based pricing across two zip codes, $36,000 for the year, and received 240 connections, $150 each. Every connection was exported from the CRM with its phone number and email, and every closed transaction from the team's transaction records with the client's. The two files were matched, and each closing was traced to the connection behind it. The average sale price was $420,000, and the team's side was assumed at 2.5%, $10,500 in gross commission per closing.

1. The calendar year undercounts by more than half. Only 4 of the year's connections had closed by 31 December. The other 5 closed over the next 18 months; the median closing came five months after the first inquiry and the longest sixteen. A renewal decision taken in January on the left-hand column would cut a program that more than doubles its money.
2. Two of the nine were already the team's. Matching on phone and email found both in the CRM before their Zillow inquiry: one a past client from 2023, one a friend's referral who browsed listings on Zillow before calling. They would very likely have bought with the team anyway, so crediting them to Zillow adds $21,000 of commission the program did not create.
3. The honest number is still a good one. Seven closings that started on Zillow returned $73,500 on $36,000: a return of 104% before the brokerage split. That is the figure to compare with the team's other sources, and the one to track from year to year.

The brokerage split matters as much as the timing. If the team keeps 70% of gross commission, the seven Zillow-sourced closings net $51,450 against $36,000 of spend, a return of 43%. The nine-closing figure would net $66,150, a return of 84%. Both are worth knowing; quote the one after the split when the question is whether the team should keep paying, because that is the money the team actually keeps.

Market-based pricing against Zillow Preferred: the break-even

A pay-at-close fee is cheaper than a fixed monthly budget when your connections close rarely, and dearer when they close often. The break-even is the annual spend divided by the fee per closing: below that number of closings Preferred costs less, above it market-based pricing does.

Using the example's figures and the hypothetical 35% fee from The Close's illustration, a Preferred closing on a $10,500 commission would cost $3,675. Divide the $36,000 annual budget by $3,675 and the break-even is 9.8 closings a year. At nine closings, Preferred would cost $33,075; at seven, $25,725. Above ten, the fixed budget is cheaper, and every additional closing widens the gap.

Three cautions keep this from being a buying guide. Preferred is invite-only, so most agents cannot choose it. The two programs do not deliver the same leads in the same volume, so the closings you would get under each are not the same closings. And the fee rate, the window and what counts as a fee-bearing transaction are set by your agreement, which is the place to look before relying on any figure here. What the table does show is that the comparison runs on your own connection-to-closing rate, and that no one can quote it to you without your data.

Who gets credit when a Zillow lead was already your client

Most buyers find their agent through people they know, and many of them browse Zillow anyway. When a past client or sphere contact inquires through Zillow, credit the relationship that came first. Otherwise Zillow's numbers absorb your referral business and your sphere looks smaller than it is.

Put those together and the overlap is predictable. Six in ten buyers come through a relationship, and most of them start looking online, often on the largest portal. Some will inquire on a listing before they think to call the agent their sister used. If your CRM records the Zillow connection as the lead source because it was the most recent touch, Zillow gets credit for a client the relationship delivered.

• Credit the earliest source you can prove. A contact who was in your CRM as a past client or a named referral before the Zillow inquiry is a sphere closing. Referral revenue covers how to record those so they are matchable.
• Match on phone and email, not on name. Portal leads arrive with typed names and phone numbers in every format. Normalize both before matching; phone normalization shows why the format alone hides half the matches.
• Collapse duplicate inquiries. The same buyer often inquires on two portals and your website in one week. Treat them as one person with several touches, as duplicate leads describes, and decide which touch gets credit by rule, not by whoever logged it.
• Keep attribution and fees separate. Who you credit in your own report does not change what your agreement says you owe. Report Zillow-sourced closings for decisions and Zillow-connected closings for reconciling invoices.

There is also value running the other way. A buyer who arrived on Zillow and had a good experience is now in your sphere, and their referrals in later years are not Zillow closings but would not exist without one. Customer lifetime value by channel covers how to see that second-order value without crediting it to the original source twice.

How to measure Zillow ROI from your own records

Export your connections and your closed transactions, match them on normalized phone and email, credit each closing to its earliest source, and report commission by connection month. Repeat quarterly and let each cohort's figure grow until it stops changing.

1. Export every lead with its source and first-contact date. From Follow Up Boss or whichever CRM receives your connections: name, phone, email, source, created date. Include sphere, past clients, open houses and other portals, not just Zillow, or there is nothing to find the overlap against. CRM export covers the columns.
2. Export closed transactions with the client's contact details. Closing date, sale price, your side's gross commission, and the buyer's or seller's phone and email from the transaction file.
3. Match closings to leads. Normalize phone numbers and emails, join the two files, and give each closing the source of its earliest matched lead. Closings with no matching lead go in Direct / Unknown rather than into the nearest channel.
4. Group Zillow-sourced closings by connection month. For each month's connections, add up the commission they have produced so far, and set it against that month's spend.
5. Review quarterly and renew on cohorts at least twelve months old. Recent cohorts are still filling; judge a zip code on the months that have had time to close, and watch the recent ones for early signs.

CloseRev does step three from two CSV files you export: closed sales and leads. It matches each closing to leads on normalized phone and email and, by default, credits the source of the earliest matched lead, so a past client who later inquired on Zillow stays with the original source; last touch is an option. Only high-confidence matches count automatically; medium- and low-confidence ones stay reviewable, and closings with no match stay in Direct / Unknown. It does not connect to Zillow or Follow Up Boss, and it does not calculate referral fees.

What to check before renewing a zip code

Before renewing or raising a zip code budget, check the closed commission from that zip's connections at least twelve months old, the share of closings that were already yours, and the cost per closing against your other sources. Connection volume alone is not a reason to renew.

• Cohort return after the split. Commission from the zip's connections, at least twelve months old, after your brokerage split, against what those months cost.
• Overlap with your sphere. The share of Zillow-connected closings whose earliest source was a relationship. A high share means you are paying for clients you already had.
• Cost per closing against other sources. Zillow's $5,143 per closing in the example is cheap next to a $9,000 open-house program and dear next to a sphere campaign costing $1,200 a closing. Customer acquisition cost by channel shows how to lay them side by side.
• Share of voice drift. If your connections fell while your spend held, someone else raised their budget in the zip. That is a price rise, and it belongs in the decision.
• Follow-up, before blaming the leads. A low connection-to-closing rate can mean slow follow-up as easily as poor leads. Missed calls and speed to lead covers the response times that decide it.

Zillow is neither a scam nor a sure thing. It is a lead source with a long tail, a fixed or percentage price, and an overlap with the business you already have. Measured by cohort, credited to the earliest source and reported after the split, it produces a number you can compare with everything else you pay for. Lead marketplaces covers the same discipline for Angi, Thumbtack and HomeAdvisor, which share Zillow's model and its measurement problems.